390 lines
8.8 KiB
JavaScript
390 lines
8.8 KiB
JavaScript
/**
|
|
xxHash implementation in pure Javascript
|
|
|
|
Copyright (C) 2013, Pierre Curto
|
|
MIT license
|
|
*/
|
|
var UINT32 = require('cuint').UINT32
|
|
|
|
/*
|
|
Merged this sequence of method calls as it speeds up
|
|
the calculations by a factor of 2
|
|
*/
|
|
// this.v1.add( other.multiply(PRIME32_2) ).rotl(13).multiply(PRIME32_1);
|
|
UINT32.prototype.xxh_update = function (low, high) {
|
|
var b00 = PRIME32_2._low
|
|
var b16 = PRIME32_2._high
|
|
|
|
var c16, c00
|
|
c00 = low * b00
|
|
c16 = c00 >>> 16
|
|
|
|
c16 += high * b00
|
|
c16 &= 0xFFFF // Not required but improves performance
|
|
c16 += low * b16
|
|
|
|
var a00 = this._low + (c00 & 0xFFFF)
|
|
var a16 = a00 >>> 16
|
|
|
|
a16 += this._high + (c16 & 0xFFFF)
|
|
|
|
var v = (a16 << 16) | (a00 & 0xFFFF)
|
|
v = (v << 13) | (v >>> 19)
|
|
|
|
a00 = v & 0xFFFF
|
|
a16 = v >>> 16
|
|
|
|
b00 = PRIME32_1._low
|
|
b16 = PRIME32_1._high
|
|
|
|
c00 = a00 * b00
|
|
c16 = c00 >>> 16
|
|
|
|
c16 += a16 * b00
|
|
c16 &= 0xFFFF // Not required but improves performance
|
|
c16 += a00 * b16
|
|
|
|
this._low = c00 & 0xFFFF
|
|
this._high = c16 & 0xFFFF
|
|
}
|
|
|
|
/*
|
|
* Constants
|
|
*/
|
|
var PRIME32_1 = UINT32( '2654435761' )
|
|
var PRIME32_2 = UINT32( '2246822519' )
|
|
var PRIME32_3 = UINT32( '3266489917' )
|
|
var PRIME32_4 = UINT32( '668265263' )
|
|
var PRIME32_5 = UINT32( '374761393' )
|
|
|
|
/**
|
|
* Convert string to proper UTF-8 array
|
|
* @param str Input string
|
|
* @returns {Uint8Array} UTF8 array is returned as uint8 array
|
|
*/
|
|
function toUTF8Array (str) {
|
|
var utf8 = []
|
|
for (var i=0, n=str.length; i < n; i++) {
|
|
var charcode = str.charCodeAt(i)
|
|
if (charcode < 0x80) utf8.push(charcode)
|
|
else if (charcode < 0x800) {
|
|
utf8.push(0xc0 | (charcode >> 6),
|
|
0x80 | (charcode & 0x3f))
|
|
}
|
|
else if (charcode < 0xd800 || charcode >= 0xe000) {
|
|
utf8.push(0xe0 | (charcode >> 12),
|
|
0x80 | ((charcode>>6) & 0x3f),
|
|
0x80 | (charcode & 0x3f))
|
|
}
|
|
// surrogate pair
|
|
else {
|
|
i++;
|
|
// UTF-16 encodes 0x10000-0x10FFFF by
|
|
// subtracting 0x10000 and splitting the
|
|
// 20 bits of 0x0-0xFFFFF into two halves
|
|
charcode = 0x10000 + (((charcode & 0x3ff)<<10)
|
|
| (str.charCodeAt(i) & 0x3ff))
|
|
utf8.push(0xf0 | (charcode >>18),
|
|
0x80 | ((charcode>>12) & 0x3f),
|
|
0x80 | ((charcode>>6) & 0x3f),
|
|
0x80 | (charcode & 0x3f))
|
|
}
|
|
}
|
|
|
|
return new Uint8Array(utf8)
|
|
}
|
|
|
|
/**
|
|
* XXH object used as a constructor or a function
|
|
* @constructor
|
|
* or
|
|
* @param {Object|String} input data
|
|
* @param {Number|UINT32} seed
|
|
* @return ThisExpression
|
|
* or
|
|
* @return {UINT32} xxHash
|
|
*/
|
|
function XXH () {
|
|
if (arguments.length == 2)
|
|
return new XXH( arguments[1] ).update( arguments[0] ).digest()
|
|
|
|
if (!(this instanceof XXH))
|
|
return new XXH( arguments[0] )
|
|
|
|
init.call(this, arguments[0])
|
|
}
|
|
|
|
/**
|
|
* Initialize the XXH instance with the given seed
|
|
* @method init
|
|
* @param {Number|Object} seed as a number or an unsigned 32 bits integer
|
|
* @return ThisExpression
|
|
*/
|
|
function init (seed) {
|
|
this.seed = seed instanceof UINT32 ? seed.clone() : UINT32(seed)
|
|
this.v1 = this.seed.clone().add(PRIME32_1).add(PRIME32_2)
|
|
this.v2 = this.seed.clone().add(PRIME32_2)
|
|
this.v3 = this.seed.clone()
|
|
this.v4 = this.seed.clone().subtract(PRIME32_1)
|
|
this.total_len = 0
|
|
this.memsize = 0
|
|
this.memory = null
|
|
|
|
return this
|
|
}
|
|
XXH.prototype.init = init
|
|
|
|
/**
|
|
* Add data to be computed for the XXH hash
|
|
* @method update
|
|
* @param {String|Buffer|ArrayBuffer} input as a string or nodejs Buffer or ArrayBuffer
|
|
* @return ThisExpression
|
|
*/
|
|
XXH.prototype.update = function (input) {
|
|
var isString = typeof input == 'string'
|
|
var isArrayBuffer
|
|
|
|
// Convert all strings to utf-8 first (issue #5)
|
|
if (isString) {
|
|
input = toUTF8Array(input)
|
|
isString = false
|
|
isArrayBuffer = true
|
|
}
|
|
|
|
if (typeof ArrayBuffer !== "undefined" && input instanceof ArrayBuffer)
|
|
{
|
|
isArrayBuffer = true
|
|
input = new Uint8Array(input);
|
|
}
|
|
|
|
var p = 0
|
|
var len = input.length
|
|
var bEnd = p + len
|
|
|
|
if (len == 0) return this
|
|
|
|
this.total_len += len
|
|
|
|
if (this.memsize == 0)
|
|
{
|
|
if (isString) {
|
|
this.memory = ''
|
|
} else if (isArrayBuffer) {
|
|
this.memory = new Uint8Array(16)
|
|
} else {
|
|
this.memory = new Buffer(16)
|
|
}
|
|
}
|
|
|
|
if (this.memsize + len < 16) // fill in tmp buffer
|
|
{
|
|
// XXH_memcpy(this.memory + this.memsize, input, len)
|
|
if (isString) {
|
|
this.memory += input
|
|
} else if (isArrayBuffer) {
|
|
this.memory.set( input.subarray(0, len), this.memsize )
|
|
} else {
|
|
input.copy( this.memory, this.memsize, 0, len )
|
|
}
|
|
|
|
this.memsize += len
|
|
return this
|
|
}
|
|
|
|
if (this.memsize > 0) // some data left from previous update
|
|
{
|
|
// XXH_memcpy(this.memory + this.memsize, input, 16-this.memsize);
|
|
if (isString) {
|
|
this.memory += input.slice(0, 16 - this.memsize)
|
|
} else if (isArrayBuffer) {
|
|
this.memory.set( input.subarray(0, 16 - this.memsize), this.memsize )
|
|
} else {
|
|
input.copy( this.memory, this.memsize, 0, 16 - this.memsize )
|
|
}
|
|
|
|
var p32 = 0
|
|
if (isString) {
|
|
this.v1.xxh_update(
|
|
(this.memory.charCodeAt(p32+1) << 8) | this.memory.charCodeAt(p32)
|
|
, (this.memory.charCodeAt(p32+3) << 8) | this.memory.charCodeAt(p32+2)
|
|
)
|
|
p32 += 4
|
|
this.v2.xxh_update(
|
|
(this.memory.charCodeAt(p32+1) << 8) | this.memory.charCodeAt(p32)
|
|
, (this.memory.charCodeAt(p32+3) << 8) | this.memory.charCodeAt(p32+2)
|
|
)
|
|
p32 += 4
|
|
this.v3.xxh_update(
|
|
(this.memory.charCodeAt(p32+1) << 8) | this.memory.charCodeAt(p32)
|
|
, (this.memory.charCodeAt(p32+3) << 8) | this.memory.charCodeAt(p32+2)
|
|
)
|
|
p32 += 4
|
|
this.v4.xxh_update(
|
|
(this.memory.charCodeAt(p32+1) << 8) | this.memory.charCodeAt(p32)
|
|
, (this.memory.charCodeAt(p32+3) << 8) | this.memory.charCodeAt(p32+2)
|
|
)
|
|
} else {
|
|
this.v1.xxh_update(
|
|
(this.memory[p32+1] << 8) | this.memory[p32]
|
|
, (this.memory[p32+3] << 8) | this.memory[p32+2]
|
|
)
|
|
p32 += 4
|
|
this.v2.xxh_update(
|
|
(this.memory[p32+1] << 8) | this.memory[p32]
|
|
, (this.memory[p32+3] << 8) | this.memory[p32+2]
|
|
)
|
|
p32 += 4
|
|
this.v3.xxh_update(
|
|
(this.memory[p32+1] << 8) | this.memory[p32]
|
|
, (this.memory[p32+3] << 8) | this.memory[p32+2]
|
|
)
|
|
p32 += 4
|
|
this.v4.xxh_update(
|
|
(this.memory[p32+1] << 8) | this.memory[p32]
|
|
, (this.memory[p32+3] << 8) | this.memory[p32+2]
|
|
)
|
|
}
|
|
|
|
p += 16 - this.memsize
|
|
this.memsize = 0
|
|
if (isString) this.memory = ''
|
|
}
|
|
|
|
if (p <= bEnd - 16)
|
|
{
|
|
var limit = bEnd - 16
|
|
|
|
do
|
|
{
|
|
if (isString) {
|
|
this.v1.xxh_update(
|
|
(input.charCodeAt(p+1) << 8) | input.charCodeAt(p)
|
|
, (input.charCodeAt(p+3) << 8) | input.charCodeAt(p+2)
|
|
)
|
|
p += 4
|
|
this.v2.xxh_update(
|
|
(input.charCodeAt(p+1) << 8) | input.charCodeAt(p)
|
|
, (input.charCodeAt(p+3) << 8) | input.charCodeAt(p+2)
|
|
)
|
|
p += 4
|
|
this.v3.xxh_update(
|
|
(input.charCodeAt(p+1) << 8) | input.charCodeAt(p)
|
|
, (input.charCodeAt(p+3) << 8) | input.charCodeAt(p+2)
|
|
)
|
|
p += 4
|
|
this.v4.xxh_update(
|
|
(input.charCodeAt(p+1) << 8) | input.charCodeAt(p)
|
|
, (input.charCodeAt(p+3) << 8) | input.charCodeAt(p+2)
|
|
)
|
|
} else {
|
|
this.v1.xxh_update(
|
|
(input[p+1] << 8) | input[p]
|
|
, (input[p+3] << 8) | input[p+2]
|
|
)
|
|
p += 4
|
|
this.v2.xxh_update(
|
|
(input[p+1] << 8) | input[p]
|
|
, (input[p+3] << 8) | input[p+2]
|
|
)
|
|
p += 4
|
|
this.v3.xxh_update(
|
|
(input[p+1] << 8) | input[p]
|
|
, (input[p+3] << 8) | input[p+2]
|
|
)
|
|
p += 4
|
|
this.v4.xxh_update(
|
|
(input[p+1] << 8) | input[p]
|
|
, (input[p+3] << 8) | input[p+2]
|
|
)
|
|
}
|
|
p += 4
|
|
} while (p <= limit)
|
|
}
|
|
|
|
if (p < bEnd)
|
|
{
|
|
// XXH_memcpy(this.memory, p, bEnd-p);
|
|
if (isString) {
|
|
this.memory += input.slice(p)
|
|
} else if (isArrayBuffer) {
|
|
this.memory.set( input.subarray(p, bEnd), this.memsize )
|
|
} else {
|
|
input.copy( this.memory, this.memsize, p, bEnd )
|
|
}
|
|
|
|
this.memsize = bEnd - p
|
|
}
|
|
|
|
return this
|
|
}
|
|
|
|
/**
|
|
* Finalize the XXH computation. The XXH instance is ready for reuse for the given seed
|
|
* @method digest
|
|
* @return {UINT32} xxHash
|
|
*/
|
|
XXH.prototype.digest = function () {
|
|
var input = this.memory
|
|
var isString = typeof input == 'string'
|
|
var p = 0
|
|
var bEnd = this.memsize
|
|
var h32, h
|
|
var u = new UINT32
|
|
|
|
if (this.total_len >= 16)
|
|
{
|
|
h32 = this.v1.rotl(1).add( this.v2.rotl(7).add( this.v3.rotl(12).add( this.v4.rotl(18) ) ) )
|
|
}
|
|
else
|
|
{
|
|
h32 = this.seed.clone().add( PRIME32_5 )
|
|
}
|
|
|
|
h32.add( u.fromNumber(this.total_len) )
|
|
|
|
while (p <= bEnd - 4)
|
|
{
|
|
if (isString) {
|
|
u.fromBits(
|
|
(input.charCodeAt(p+1) << 8) | input.charCodeAt(p)
|
|
, (input.charCodeAt(p+3) << 8) | input.charCodeAt(p+2)
|
|
)
|
|
} else {
|
|
u.fromBits(
|
|
(input[p+1] << 8) | input[p]
|
|
, (input[p+3] << 8) | input[p+2]
|
|
)
|
|
}
|
|
h32
|
|
.add( u.multiply(PRIME32_3) )
|
|
.rotl(17)
|
|
.multiply( PRIME32_4 )
|
|
p += 4
|
|
}
|
|
|
|
while (p < bEnd)
|
|
{
|
|
u.fromBits( isString ? input.charCodeAt(p++) : input[p++], 0 )
|
|
h32
|
|
.add( u.multiply(PRIME32_5) )
|
|
.rotl(11)
|
|
.multiply(PRIME32_1)
|
|
}
|
|
|
|
h = h32.clone().shiftRight(15)
|
|
h32.xor(h).multiply(PRIME32_2)
|
|
|
|
h = h32.clone().shiftRight(13)
|
|
h32.xor(h).multiply(PRIME32_3)
|
|
|
|
h = h32.clone().shiftRight(16)
|
|
h32.xor(h)
|
|
|
|
// Reset the state
|
|
this.init( this.seed )
|
|
|
|
return h32
|
|
}
|
|
|
|
module.exports = XXH
|